Molecular recognition of histidine-tagged molecules by metal-chelating lipids monitored by fluorescence energy transfer and correlation spectroscopy

Article abstract of DOI:10.1021/ja9735620

Complex binding of proteins by metal-chelating lipids via surface- exposed or protein-engineered histidines provides an universal and powerful concept for the orientation and two-dimensional crystallization of proteins at self-organized interfaces. To dem

Full text of DOI:10.1021/ja9735620

J. Am. Chem. Soc. 1998, 120, 2753-2763
2753
Molecular Recognition of Histidine-Tagged Molecules by
Metal-Chelating Lipids Monitored by Fluorescence Energy Transfer
and Correlation Spectroscopy§
Ingmar T. Dorn,† Klaus R. Neumaier,† and Robert Tampe´*,†,‡
Contribution from Lehrstuhl fu¨r Biophysik, Technische UniVersita¨t Mu¨nchen, James-Franck-Strasse,
D-85747 Garching, Germany, and Max-Planck-Institut fu¨r Biochemie, Am Klopferspitz 18a,
D-82152 Martinsried, Germany
ReceiVed October 13, 1997
Abstract: Complex binding of proteins by metal-chelating lipids via surface-exposed or protein-engineered
histidines provides an universal and powerful concept for the orientation and two-dimensional crystallization
of proteins at self-organized interfaces. To demonstrate pair formation between individual histidine-tagged
molecules and chelator lipids on the molecular level, we have synthesized novel lipids bearing both a Ni-
NTA chelator and a fluorescent group. These lipids serve as spectroscopic probes to visualize directly the
molecular recognition of fluorescence-labeled histidine-tagged peptides by metal-chelating lipids using
fluorescence resonance energy transfer (FRET). The molecular docking to chelator lipids assembled in mono-
or bilayers is highly specific, revealing only 3% unspecific adsorption and a binding constant of 3 µM. The
affinity constant was confirmed by fluorescence correlation spectroscopy (FCS) on single molecules, where
the ratio of lipid-bound to free was analyzed by their intrinsically different diffusion times passing through a
confocal volume of about 1 fL. By using a model peptide most of the electrostatic and steric contribution to
the binding process can be neglected. Therefore, the affinity constant can serve as a standard value for the
binding of histidine-tagged proteins to chelator lipid interfaces.
Introduction
assembled monolayers of alkyl silanes5 or thiols6,7 on glass or
gold, respectively, or (iii) by self-assembled lipid layers. Unlike
other thin films, lipid mono- or bilayers can be deposited on
nearly every surface by various techniques8,9 providing bio-
compatibility, lateral mobility, and two-dimensional patterning.
Functional units of biomembranes such as channels, transporters,
or other membrane proteins can be reconstituted into vesicles
and immobilized at the surface by vesicle fusion.10-12 Due to
their dynamic properties, lipids can be organized in two
dimensions by phase segregation,13 electrical fields,14 and
microfabricated barriers,15 allowing the generation of structured
biofunctional interfaces. Furthermore, lateral diffusion is still
possible in supported membranes which is important to mimic
processes at or within biological membranes.
Biofunctionalization of interfaces plays an important role in
bioanalytical, biochemical, biophysical, and structural studies.
Such two-dimensional systems can be studied with well-
developed surface-sensitive methods. Therefore, solid surfaces
have to be coated by thin films that are functional and
biocompatible to avoid distortion of the studied systems caused
by unspecific interactions. This surface coating can be achieved
by several materials: (i) by adsorbed polymers,1-4 (ii) by self-
* Corresponding author: Robert Tampe´, Max-Planck-Institut fu¨r Bio-
chemie, Am Klopferspitz 18a, D-82152 Martinsried, Germany. Tele-
phone: +49-89-8578-2646. Fax: +49-89-8578-2641. E-mail: tampe@
biochem.mpg.de.
† Technische Universita¨t Mu¨nchen.
‡ Max-Planck-Institut fu¨r Biochemie.
(5) Brzoska, J. B.; Azousz, I. B.; Rondelez, F. Langmuir 1994, 10, 4367-
4373.
(6) Dubois, L. H.; Nuzzo, R. G. Annu. ReV. Phys. Chem. 1992, 43, 437-
463.
(7) Whitesides, G. M.; Gorman, C. G. Handbook of Surface Imaging
and Visualization; CRC Press: Boca Raton, FL, 1995; pp 713-733.
(8) Sackmann, E. Science 1996, 271, 43-48.
(9) Tampe´, R.; Dietrich, C.; Elender, G.; Gritsch, S.; Schmitt, L.
Nanofabrication and Biosystems: Integrating Materials Science, Enge-
neering, and Biology; Cambridge University Press: London, 1996; pp 201-
221.
(10) Brian, A. A.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1984,
81, 6159-6163.
(11) Chan, P.; Lawrence, M. B.; Dustin, M. L.; Ferguson, L. M.; Golan,
D. E.; Springer, T. A. J. Cell Biol. 1991, 10, 245-255.
(12) Salafsky, J.; Groves, J. T.; Boxer, S. G. Biochemistry 1996, 35,
14773-14781.
§ Abbreviations: AcOH, acetic acid; amu, atom mass unit; au, arbitrary
units; Boc, tert-butyloxycarbonyl; cpm, counts per molecule; DMF, N,N-
dimethylformamide; DCC, N,N′-dicyclohexylcarbodiimide; DMSO, di-
methyl sulfoxide; DODA, dioctadecylamine; EDTA, ethylenediaminetet-
raacetic acid; EI, electron impact; FAB, fast-atom bombardment; FCS,
fluorescence correlation spectroscopy; FRET, fluorescence resonance energy
transfer; HEPES, N-2-hydroxyethylpiperazine-N′-ethanesulfonic acid; IDA,
iminodiacetic acid; MeOH, methanol; MS, mass spectrometry; m/z, mass
per charge; NBD, 7-nitro-2,1,3-benzoxadiazol-4-yl; NHS, N-hydroxysuc-
cinimide; Ni-NTA, nickel complexed by NTA; NTA, nitrilotriacetic acid;
SOPC, 1-stearoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine; SPR, surface
plasmon resonance; TFA, trifluoroacetic acid; THF, tetrahydrofuran; TLC,
thin-layer chromatography; Z, benzyloxycarbonyl.
(1) Lo¨fås, S.; Johnsson, B. J. Chem. Soc., Chem. Commun. 1990, 21,
1526-1528.
(2) Ku¨hner, M.; Tampe´, R.; Sackmann, E. Biophys. J. 1994, 67, 217-
226.
(3) Beyer, D.; Knoll, W.; Ringsdorf, H.; Elender, G.; Sackmann, E. Thin
Solid Films 1996, 285, 825-828.
(4) Elender, G.; Ku¨hner, M.; Sackmann, E. Biosens. Bioelect. 1996, 11,
565-577.
(13) McConnell, H. M. Annu. ReV. Phys. Chem. 1991, 42, 171-95.
(14) Dietrich, C.; Tampe´, R. Biochim. Biophys. Acta 1995, 1238, 183-
191.
(15) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651-
653.
S0002-7863(97)03562-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/13/1998

2754 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Dorn et al.
Functional and structural studies often require the oriented
and functional immobilization of proteins which can be achieved
in several ways. Physisorption of a protein to a surface via
van der Waals, electrostatic, or entropic interactions are highly
sensitive toward ionic strength, pH, or temperature. This can
lead to multilayer adsorption or a loss of orientation and activity
of the protein. Covalent binding of the protein via surface-
accessible amino acid residues to a reactive surface often lacks
regiospecificity and therefore orientation of the immobilized
protein. Additionally, the reactive site of a protein can also be
blocked by the immobilization procedure which leads to a
reduced activity of the protein.16 Protein binding via its natural
ligand17,18 is only possible if the ligands can be fixed at the
surface and if the process of ligand binding is not the process
desired to study. The very affine binding of biotin to strepta-
vidin can also be used for protein immobilization.17 However,
this method requires a rather unspecific chemical biotinylation
of the protein. Protein-engineered affinity tags mimicking the
biotin moiety overcome this problem. However, they have a
drastically reduced affinity to streptavidin19,20 and an im-
mobilized streptavidin layer as a docking interface is still needed.
Immobilized metal ion affinity chromatography (IMAC)21,22 is
a versatile and powerful method for protein purification and
characterization. Here, fusion proteins are expressed with a
short affinity sequence of five or six histidines (histidine-tag)
that binds to metal ion complexes such as nitrilotriacetic acid
(NTA) or iminodiacetic acid (IDA). In most cases, the
functionality of the protein is preserved. Additionally, the
binding is reversible and dissociation can be induced at low
pH or by addition of a competitor either for the protein
(histidine, imidazole) or the metal ion (EDTA).
segregated chelator lipids. The histidine-tagged protein was
found to be fully functional with respect to its specific DNA
recognition. Two-dimensional crystallization of proteins at
chelator lipid monolayers was reported.33-36 Metal-chelating
surfaces can be produced either by coating with chelator lipids30
or by modification of different surfaces with NTA.16,37,38
Although specific binding to self-assembled metal-chelating
interfaces has been demonstrated, it was not possible to
distinguish between adsorption to the interface and direct pair
formation between the chelator and the biomolecule.
In this report, we describe the synthesis of fluorescent chelator
lipids which serve as spectroscopic probes to follow the binding
process. Due to a distance dependency of fluorescence reso-
nance energy transfer (FRET) in the nanometer range, pair
formation between an acceptor-labeled histidine-tagged molecule
and a donor-labeled chelator lipid was proven. Molecular
recognition was followed at lipid monolayers at the air-water
interface as well as at lipid bilayers in vesicle solution.
Additionally, specific binding of single histidine-tagged mol-
ecules to chelator lipid containing vesicles could be demon-
strated by fluorescence correlation spectroscopy (FCS). Due
to their intrinsically different diffusion times for passing through
a confocal volume, the ratio of free and lipid-bound molecules
could be analyzed by the autocorrelation function of the time-
dependent fluorescence signal. The quantitative and molecular
analysis of this highly specific and oriented docking process
provides deeper insights into this versatile and powerful method
for reversible immobilization and orientation of proteins at lipid
interfaces.
Materials and Methods
We as well as Arnold and colleagues combined the IMAC
technique with the unique and fascinating properties of self-
organizing systems by introducing the concept of chelator
lipids.23,24 The metal ion and ligand binding properties of the
chelator lipid interfaces as well as their dependency on ionic
strength, pH, and temperature was examined in detail.25,26
Specific docking of peptides27,28 and proteins to chelator lipids
via histidine tag29-31 or surface histidines24 could be demon-
strated. A designed model peptide32 as well as a DNA binding
protein29 could be organized in two dimensions by phase-
Materials. The chemicals used were 1-stearoyl-2-oleoyl-sn-glycero-
3-phosphatidylcholine (Avanti Polar Lipids, Birmingham, AL), Nꢀ-
benzyloxycarbonyl-L-lysine tert-butyl ester hydrochloride (Bachem,
Heidelberg, Germany), Pd/C (5%), fluorescamine, rhodamine 6G,
succinic anhydride (Aldrich, Steinheim, Germany), and 5-(and-6)-
carboxytetramethylrhodamine succinimidyl ester (Molecular Probes,
Eugene, OR). Thin-layer chromatography (TLC) plates (silica gel 60
F254), silica gel (40-63 µM, 230-400 mesh), bromocresol green, and
ninhydrin were purchased from Merck (Darmstadt, Germany). All other
chemicals and solvents were ordered from Fluka (Neu-Ulm, Germany)
and were reagent p.a. grade. Solvent ratios are given in volume/volume
if not otherwise stated. The histidine peptide H-GSGSGSGSGSHH-
HHHH-OH was synthesized by a solid-phase technique using fluore-
nylmethoxycarbonyl chemistry. The N-terminus was labeled with
tetramethylrhodamine succinimidyl ester in DMSO. The fluorescence-
labeled peptide was isolated by reversed-phase HPLC (C18-column)
with an gradient from 20 to 80% acetonitrile (0.1% TFA) in water
(0.1% TFA). The identity of the labeled peptide was verified by mass
spectrometry (electrospray ionization, M + H+ ) 1977 m/z).
(16) Gershon, P. D.; Khilko, S. J. Immunol. Methods. 1995, 183, 65-
76.
(17) Ahlers, M.; Mu¨ller, W.; Reichert, A.; Ringsdorf, H.; Venzmer, J.
Angew. Chem. 1990, 102, 1310-1327.
(18) Madoz, J.; Kuznetzov, B. A.; Medrano, F. J.; Garcia, J. L.;
Fernandez, V. M. J. Am. Chem. Soc. 1997, 119, 1043-1051.
(19) Schmidt, T. G. M.; Koepke, J.; Frank, R.; Skerra, A. J. Mol. Biol.
1996, 255, 753-766.
(20) Hengsakul, M.; Cass, A. E. G. J. Mol. Biol. 1997, 266, 621-632.
(21) Porath, J.; Carlsson, J.; Olsson, I.; Belfrage, G. Nature 1975, 258,
598-599.
Analytical Methods. Reactions were monitored by TLC and
functional groups were visualized by UV absorbance (Z group),
fluorescamine (primary amines), ninhydrin (amines), cobalt isothiocy-
(22) Hochuli, E.; Bannwarth, W.; Do¨beli, H.; Gentz, R.; Stu¨ber, D. Bio/
Technology 1988, 6, 1321-1325.
(23) Schmitt, L.; Dietrich, C.; Tampe´, R. J. Am. Chem. Soc. 1994, 116,
8485-8491.
(32) Dietrich, C.; Schmitt, L.; Tampe´, R. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 9014-9018.
(24) Shnek, D. R.; Pack, D. W.; Sasaki, D. Y.; Arnold, F. H. Langmuir
1994, 10, 2382-2388.
(33) Kubalek, E. W.; Le Grice, F. J.; Brown, P. O. J. Struct. Biol. 1994,
113, 117-123.
(25) Dietrich, C. Ph.D. Thesis, TU-Mu¨nchen 1995.
(26) Pack, D. W.; Arnold, F. H. Chem. Phys. Lipid. 1997, 86, 135-
152.
(34) Frey, W.; Schief, W. R.; Pack, D. W.; Chen, C. T.; Chilkoti, A.;
Stayton, P.; Vogel, V.; Arnold, F. H. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 4937-4941.
(27) Gritsch, S.; Neumaier, K.; Schmitt, L.; Tampe´, R. Biosens. Bioelect.
1995, 10, 805-812.
(35) Pack, D. W.; Chen, G.; Maloney, K. M.; Chen, C.; Arnold, F. H. J.
Am. Chem. Soc. 1997, 119, 2479-2487.
(28) Maloney, K. M.; Shnek, D. R.; Sasaki, D. Y.; Arnold, F. H. Chem.
Biol. 1996, 3, 185-192.
(36) Barklis, E.; McDermott, J.; Wilkens, S.; Schabtach, E.; Schmid,
M. F.; Fuller, S.; Karanjia, S.; Love, Z.; Jones, R.; Rui, Y.; Zhao, X.;
Thompson, D. EMBO J. 1997, 16, 1199-1213.
(29) Dietrich, C.; Boscheinen, O.; Scharf, K. D.; Schmitt, L.; Tampe´,
R. Biochemistry 1996, 35, 1100-1105.
(30) Schmitt, L.; Bohanon, T. M.; Denzinger, S.; Ringsdorf, H.; Tampe´,
R. Angew. Chem., Int. Ed. Engl. 1996, 35, 317-320.
(31) Ng, K.; Pack, D. W.; Sasaki, D. Y.; Arnold, F. H. Langmuir 1995,
11, 4048-4055.
(37) O’Shannessy, D. J.; O’Donnell, K. C.; Martin, J.; Brigham-Burke,
M. Anal. Biochem. 1995, 229, 119-124.
(38) Sigal, G. B.; Bamdad, C.; Barberis, A.; Strominger, J.; Whitesides,
G. M. Anal. Chem. 1996, 68, 490-497.

Molecular Recognition of Histidine-Tagged Molecules
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2755
anate (amines, amides), and bromocresol green (acids and bases). 1H
NMR (500 MHz) and 13C NMR (100 MHz) spectra were recorded on
Bruker AM 500 and AM 400, respectively. Chemical shifts (δ) are
given in parts per million relative to the CDCl3 signal. Mass
spectrometry was performed at a Finnigan MAT (Forster City, CA)
either in the electron impact (EI), in the electrospray, or in the fast-
atom bombardment (FAB) ionization mode.
In eqs 1 and 2, N is the number of molecules in the confocal volume,
τ the diffusion time through the confocal volume (in eq 2, of component
1 or 2, respectively), S the structural parameter, TA the fraction of
molecules in the triplet state, t0 the time constant of the triplet state,
and y the fraction of molecules with a diffusion time τ2.
Diffusion coefficients D were calculated by eq 3,
2
Film Balance Measurements. Epifluorescence film balance mea-
surements were performed on a self-built instrument as previously
described.32 The trough (Teflon) was 220 mm × 30 mm and carried
a subphase volume of 25 mL. The surface tension was measured by
a Wilhelmy system, and the temperature was controlled by peltier
elements. The speed of the movable barrier and the lateral pressure
were computer-controlled. The film balance was equipped with an
epifluorescence microscope. Fluorescence was excited by a HBO 50
lamp (Zeiss, Kochel, Germany). The fluorescence filter set (Zeiss)
consisted of an excitation filter (450-490 nm), a dichroitic mirror (510
nm), and an emission filter (long pass >520 nm). The fluorescence
microscope was interfaced to a monochromator via a glass fiber
technique. Intensities were measured by a photomultiplier working in
the single photon counting mode. For all experiments, degassed and
sterile filtered HEPES buffers (10 mM HEPES, 150 mM NaCl, pH
7.5) were used if not stated otherwise. The NTA lipids were preloaded
with nickel ions in organic solution by adding equimolar amounts of a
methanol solution of NiCl2‚6H2O. In experiments in which the buffer
contained EDTA, the unloaded NTA lipid was used. For the binding
experiments, the lipid monolayer was compressed to a surface pressure
of 20 mN/m at a temperature of 22 °C. The histidine peptide was
injected into the subphase while the area was kept constant. The spectra
were measured in the sample/reference (s/r) mode and baseline
corrected. The baseline was recorded prior to each experiment. To
reach equilibrium the spectra were recorded 2 or 4 h after addition of
the histidine peptide or EDTA, respectively.
D ) ω1 /4τ
(3)
where ω1 corresponds to the radius of the elliptical confocal volume
perpendicular to the optical axis and τ to the diffusion time through
the confocal volume.
Experimental Section
Synthesis of the NBD-Labeled Chelator Lipid. Synthesis of
N-(tert-Butyloxycarbonyl)-12-aminododecanoic Acid (2). 12-Amino-
dodecanoic acid (1) (430 mg, 2 mmol) was suspended in 20 mL of
absolute methanol and 400 µL of triethylamine. A solution of 872 mg
(4 mmol) of di-tert-butyl dicarbonate in 10 mL of absolute methanol
was added. After being stirred for 1 h at 60 °C, the reaction mixture
cleared up and the CO2 production stopped. The solvent was removed
in a vacuum, and the remaining oil was dissolved in 20 mL of
chloroform and extracted twice with hydrochloric acid (pH 2.5). The
organic phase was dried over anhydrous sodium sulfate, and the solvent
as well as the tert-butyl alcohol was removed in a vacuum. Yield:
605 mg (1.92 mmol) of 2; 96%. TLC: Rf ) 0.61 in CHCl3/MeOH
(15:1). 1H NMR (400 MHz, CDCl3): δ ) 1.20 (m, 16 H, Boc-NHCH2-
CH2(CH2)8CH2COOH), δ ) 1.36 (s, 9 H, (H3C)3COCONH-), δ )
1.55 (q, 2 H, Boc-NHCH2CH2(CH2)8CH2COOH), δ ) 2.26 (t, 2 H,
Boc-NHCH2CH2(CH2)8CH2COOH), δ ) 3.02 (bm, 2 H, Boc-NHCH2-
CH2(CH2)8)CH2COOH), δ ) 4.49 (bs, 1 H, Boc-NHCH2CH2(CH2)8-
CH2COOH). 13C NMR (100 MHz, CDCl3): δ ) 25.37, 27.45, 29.12,
29.71-30.10, 30.69 ((H3C)3COCONHCH2CH2(CH2)8CH2COOH), δ
) 34.72 (CH2COOH), δ ) 41.34 (CONHCH2), δ ) 79.77 ((H3C)3-
COCO), δ ) 156.76 ((H3C)3COCONH), δ ) 179.94 (CH2COOH).
MS (EI, C17H33NO4): M+ ) 315 m/z.
Synthesis of N-(N′-(tert-Butyloxycarbonyl)-12-aminododecanoyl)-
octadecylamide (3). After dissolving 605 mg (1.92 mmol) of
compound 2 and 845 mg (3.84 mmol) of 2,2′-dipyridyl disulfide in 25
mL of absolute chloroform, 516 mg (1.92 mmol) of octadecylamine,
1006 mg (3.84 mmol) of triphenylphosphine, and 300 µL of triethyl-
amine were added in 25 mL of absolute chloroform. After the mixture
was stirred for 4 h at room temperature, the solvent was evaporated
and the product was recrystallized from acetone. Yield: 1000 mg of
3, 92%. TLC: Rf ) 0.84 in CHCl3/MeOH (50:1). 1H NMR (400 MHz,
CDCl3): δ ) 0.87 (t, 3H, H3C-), δ ) 1.25 (m, 46 H, H3C(CH2)15-
CH2-, Boc-NH(CH2)2(CH2)8CH2-), δ ) 1.43 (m, 11 H, (H3C)3-
CCONHCH2CH2-), δ ) 1.60 (q, 2 H, H3C(CH2)15CH2CH2NHCO-),
δ ) 2.14 (dd, 2 H, H3C(CH2)17NHCOCH2-), δ ) 3.09 (dd, 2H, Boc-
NHCH2-), δ ) 3.22 (dd, 2 H, H3C(CH2)16CH2NHCO-), δ ) 4.51
(bs, 1 H, Boc-NH-), δ ) 5.48 (bs, 1 H, H3C(CH2)17NHCO-). 13C
NMR (100 MHz, CDCl3): δ ) 14.77 (H3C-), δ ) 23.35, 26.49, 27.44,
27.60, 29.10, 29.97-30.25, 30.35, 30.73, 32.59 (H3C(CH2)16CH2-
NHCOCH2(CH2)9CH2NHCOOC(CH3)3), δ ) 40.15 (H3C(CH2)16CH2-
NHCO-), δ ) 41.30 (Boc-NHCH2-), δ ) 79.64 ((H3C)3COCO-),
δ ) 156.66 ((H3C)3COCO-), δ ) 173.68 (H3C(CH2)17NHCO-). MS
(EI, C35H70N2O3): M+ ) 566 m/z.
Vesicle Preparation. Appropriate molar ratios of lipid were mixed
in chloroform/methanol (6:1), and the organic solvent was removed in
a vacuum. The dry lipid was swollen in HEPES buffer to a final
concentration of 6.7 mM at room temperature for 2 h and then extruded
21 times through 100-nm filters (LiposoFast-Basic, Avestin, Ottawa,
Canada). The lipid concentration was measured by UV absorption at
490 nm according to the absorption maximum of NBD.
Fluorescence Spectroscopy in Vesicle Solution. Fluorescence
spectra were recorded at a donor concentration of 3.25 µM and at
acceptor concentrations of 0.5, 1, 2.5, 5, and 10 µM on a Fluorolog-2
(Spex Industries Inc., Edison, NJ) with an excitation at 450 nm to
minimize direct excitation of the acceptor. We scanned the emission
from 490 to 620 nm in 2-nm steps with an integration time of 1 s per
step. All spectra were recorded at room temperature in the s/r mode.
Fluorescence Correlation Spectroscopy. FCS measurements were
performed on a Zeiss ConfoCor (Carl Zeiss Jena, Jena, Germany)
equipped with a Nd:YAG laser (10 mW, excitation at 532 nm). The
instrument was controlled by FCS Access software, and the autocor-
relation functions were analyzed with FCS Access Fit software.39 The
diameter of the pinhole was adjusted to 50 µM and the geometry of
the confocal element was determined by a one-component fit (eq 1) to
the correlation curve of rhodamine 6G. To determine an affinity
constant, we titrated 15 nM rhodamine-labeled histidine peptide with
an increasing concentration of SOPC vesicles doped with 3 mol %
Ni-NTA lipid. Each point on the titration curve corresponds to the
average of six measurements. The time-dependent fluorescence signal
was collected over a period of 60 s. The autocorrelation curves were
fitted by a two-component model (eq 2) as described.39
Synthesis of N-(N′-Methyl-12-aminododecyl)octadecylamine (4).
A LiAlH4-THF solution (44.2 mL, 0.2 M) was added drop by drop at
0 °C to a solution of 1000 mg (1.77 mmol) of compound 3 in 200 mL
of THF (freshly distilled over CaH2) under an inert gas atmosphere.
The reaction mixture was refluxed for 3 h and then cooled to 0 °C.
The excess of LiAlH4 was quenched by adding 336 µL of water and
336 µL of NaOH (15% in water, w/w) and 1 mL of water. The
precipitated Al(OH)3 is filtered off and washed three times with hot
chloroform. The combined organic phases were dried over anhydrous
sodium sulfate, and the solvent was removed in a vacuum. The product
was recrystallized from THF. Yield: 734 mg (1.58 mmol) of 4, 89%.
1 - TA(1 - e-t/t
N[1 - TA]
)
0
1
1
G(t) ) 1 +
(1)
2
1/2
1 + (t/τ)
[1 + S (t/τ)]
1 - TA(1 - e-t/t
)
0
G(t) ) 1 +
×
N[1 - TA]
1 - y
1 + (t/τ )
1
y
1
+
(2)
(39) EVOTEC FCS Access Fit Software; EVOTEC: Hamburg, Germany,
1996.
2
1/2
2
1/2
1 + (t/τ )
[
]
[1 + S (t/τ1)]
[1 + S (t/τ2)]
1
2

2756 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Dorn et al.
TLC: Rf ) 0.54 in CHCl3/MeOH/AcOH (40:8:1). 1H NMR (400 MHz,
CDCl3): δ ) 0.87 (t, 2 H, H3C(CH2)17NH-), δ ) 1.18 (s, 1 H, H3C-
(CH2)17NH-), δ ) 1.27 (m, 46 H, H3C(CH2)15CH2CH2NHCH2CH2-
(CH2)8-), δ ) 1.47 (m, 6 H, H3C(CH2)15CH2CH2NHCH2CH2-
(CH2)8CH2-), δ ) 2.43 (s, 3 H, H3CNH(CH2)12NH-), δ ) 2.55 (t, 4
H, -CH2NHCH2-), δ ) 2.58 (t, 2 H, H3CNHCH2-). 13C NMR (100
MHz, CDCl3): δ ) 14.10 (H3C-), δ ) 22.69, 27.36, 27.45, 29.36,
29.60-29.68, 29.97, 30.24, 31.92 (H3C(CH2)16)CH2NHCH2(CH2)10CH2-
NHCH3), δ ) 36.58 (H3CNH(CH2)12NH-), δ ) 50.19 (-CH2-
NHCH2-), δ ) 52.26 (H3CNHCH2-). MS (FAB, C31H66N2): M +
H+ ) 467 m/z.
Synthesis of N-(N′-Methyl-N′-(tert-butyloxycarbonyl)-12-amino-
dodecyl)octadecylamine (5). Di-tert-butyl dicarbonate (349 mg, 1.6
mmol) in 25 mL of absolute chloroform/methanol (9:1) was added to
a solution of 734 mg (1.58 mmol) of compound 4 and 700 µL of
triethylamine in 50 mL of absolute chloroform/methanol (9:1) at -20
°C over a period of 2 h. Thereafter the reaction was stopped by adding
50 mL of cold hydrochloric acid (pH 2.5). The cold bath was removed,
and the reaction mixture was stirred for an additional 10 min at room
temperature. After phase separation, the organic phase was extracted
with 1 N NaOH and dried over anhydrous sodium sulfate, and the
solvent was removed in a vacuum. The product was purified by silica
gel column chromatography using CHCl3/MeOH (15:1) as the eluant.
Yield: 715 mg (1.26 mmol) of 5, 80%. TLC: Rf ) 0.48 in CHCl3/
MeOH (15:1). 1H NMR (400 MHz, CDCl3): δ ) 0.88 (t, 3 H,
H3C(CH2)17NH-), δ ) 1.11 (s, 1 H, -CH2NHCH2-), δ ) 1.25 (m,
46 H, H3C(CH2)15CH2CH2NHCH2CH2(CH2)8CH2-), δ ) 1.45 (s, 9H,
(H3C)3COCON(CH3)-), δ ) 1.49 (q, 2 H, Boc-N(CH3)CH2CH2-), δ
) 1.65 (q, 4 H, -CH2CH2NHCH2CH2-), δ ) 2.74 (bt, 4 H, CH2CH2-
NHCH2CH2-), δ ) 2.83 (s, 3 H, Boc-N(CH3)-), δ ) 3.18 (bdd, 2 H,
Boc-N(CH3)CH2-). 13C NMR (100 MHz, CDCl3): δ ) 14.80
(H3C-), δ ) 23.39, 27.42, 27.96, 28.52, 29.18, 29.34, 30.12-30.38,
32.62 (H3C(CH2)16CH2NHCH2(CH2)10-), δ ) 34.73 (Boc-N(CH3)-
) 2.64 (d, 3 H, H3C+NH2CH2CH2-), δ ) 2.90 (dt, 2 H, H3C+NH2CH2-
CH2-), δ ) 3.24 (2 dd, 4 H, -CH2N(CO(CH2)2COOH)CH2-), δ )
7.70-8.40 (b, H3C+NH2CH2-). 13C NMR (100 MHz, CDCl3): δ )
14.77 (H3C-), δ ) 23.34, 26.98, 27.13, 27.41, 27.62, 27.78, 28.20,
28.40, 29.13, 29.26, 29.39, 29.55, 29.64, 30.02, 30.13-30.35, 31.42,
31.50, 32.58 (H3C(CH2)16CH2N(CO(CH2)2COOH)CH2(CH2)10-), δ )
33.61 (H3C+NH2CH2-), δ ) 46.95 (-CH2N(CO(CH2)2COOH)-
CH2-), δ ) 48.68 (-CH2N(CO(CH2)2COOH)CH2-), δ ) 50.09
(H3C+NH2CH2-), δ ) 172.73 (N(CO(CH2)2COOH)), δ ) 177.66
(N(CO(CH2)2COOH). MS (FAB, C35H70N2O3): M + H+ ) 667 m/z.
Synthesis of N-Succinyl-N-(N′-methyl-N′-(7-nitro-2,1,3-benzoxa-
diazol-4-yl)-12-aminododecyl)octadecylamine (8). Compound 7 (617
mg, 1.09 mmol) and 300 µL of triethylamine were dissolved in 40 mL
of chloroform, and 239 mg (1.20 mmol) of 7-nitro-2,1,3-benzoxadiazol-
4-chloride was added under the exclusion of light. After 12 h of stirring
at room temperature, the organic phase was extracted three times with
1 N hydrochloric acid and dried over anhydrous sodium sulfate, and
the solvent was removed in a vacuum. The product 8 was recrystallized
from acetone. Yield: 715 mg (0.98 mmol) of 8, 90%. TLC: Rf )
0.45 in CHCl3/MeOH (15:1). 1H NMR (400 MHz, CDCl3): δ ) 0.87
(t, 3 H, H3C(CH2)17N), δ ) 1.26 (m, 46 H, H3C(CH2)15CH2-
CH2N(CO(CH2)2COOH)CH2CH2(CH2)8-), δ ) 1.53 (m, 4 H, -CH2-
CH2N(CO(CH2)2COOH)CH2CH2-), δ ) 1.75 (q, 2 H, NBD-N(CH3)-
CH2CH2-), δ ) 2.67 (m, 4 H, -CH2N(CO(CH2)2COOH)CH2-), δ
) 3.23, 3.30 (2 dd, 2*2 H, -CH2N(CO(CH2)2COOH)CH2), δ ) 3.48
(m, 3 H, NBD-N(CH3)CH2-), δ ) 4.05 (m, 2 H, NBD-N(CH3)-
CH2-), δ ) 6.07 (d, 1 H, H-6), δ ) 8.41 (d, 1 H, H-5). 13C NMR
(100 MHz, CDCl3): δ ) 14.77 (H3C-), δ ) 23.35, 27.30, 27.56, 27.64,
28.29, 28.70, 29.49, 29.91, 30.01, 30.12-30.35, 31.04, 32.59
(H3C(CH2)15CH2N(CO(CH2)2COOH)CH2(CH2)10-), δ ) 42.13 (NBD-
N(CH3)CH2-), δ ) 47.31 (-CH2N(CO(CH2)2COOH)CH2-), δ )
48.97 (-CH2N(CO(CH2)2COOH)CH2-), δ ) 56.66 (NBD-N(CH3)-
CH2-), δ ) 101.61 (C-6), δ ) 122.76 (C-5), δ ) 136.06 (C-1), δ )
145.19 (C-2), δ ) 145.55 (C-3), δ ) 146.13 (C-4), δ ) 172.78
CH2-),
δ ) 49.44, 49.98 (-CH2NHCH2-), δ ) 79.71
((H3CCCON(CH3)-), δ ) 156.70 ((H3C)3COCON(CH3)-). MS
(-CH2N(CO(CH2)2COOH)CH2),
δ
)
175.99 (-CH2N(CO-
(FAB, C36H74N2O2): M + H+ ) 567 m/z.
(CH2)2COOH)CH2-). MS (FAB, C41H71N5O6): M + H+ ) 730 m/z.
Synthesis of Nr,Nr-Bis[(tert-butyloxycarbonyl)methyl]-NE-ben-
zyloxycarbonyl-L-lysine tert-Butyl Ester (10). A solution of 535 mg
(1.65 mmol) Nꢀ-benzyloxycarbonyl-L-lysine tert-butyl ester hydrochlo-
ride, 1.33 mL of triethylamine, and 3.2 g (16.5 mmol) of bromoacetic
acid tert-butyl ester was stirred in 20 mL of DMF for 4 days at 50 °C.
The solvent and the excess of bromoacetic acid tert-butyl ester was
removed in a vacuum, and the remaining oil was extracted six times
with hexane. The combined organic phases were collected and the
solvent was removed in a vacuum. Yield: 614 mg (1.09 mmol) of
10, 66%. TLC: Rf ) 0.70 in CHCl3/MeOH (100:1). 1H NMR (400
MHz, CDCl3): δ ) 1.42 (s, 18 H, ((H3C)3COCOCH2)2N-), δ ) 1.45
(s, 9H, (H3C)3COCOCH2CH-) δ ) 1.53 (m, 4 H, Z-NHCH2-
(CH2)2-), δ ) 1.64 (m, 2 H, Z-NH(CH2)3CH2-), δ ) 3.19 (q, 2 H,
Z-NHCH2-), δ ) 3.31 (t, 1 H, Z-NH(CH2)4CH-), δ ) 3.46 (dd, 4 H,
((H3C)3COCOCH2)2N-), δ ) 5.08 (s, 2 H, C6H5CH2OCONH-), δ )
7.32 (m, 5 H, C6H5CH2OCONH-). 13C NMR (100 MHz, CDCl3): δ
) 23.68, 28.77, 28.88-29.90, 30.74 (Z-NHCH2(CH2)3-, ((CH3)3-
COCO)3), δ ) 41.47 (Z-NHCH2-), δ ) 54.56 (-N(CH2)COOC(CH3)2),
δ ) 65.83 (-CHN(CH2COOC(CH3)3)2), δ ) 67.10 (C6H5CH2-
OCONH-), δ ) 81.44 (-N(CH2COOC(CH3)3)2), δ ) 81.85 ((H3C)3-
COCOCH-), δ ) 128.60-128.72, 129.10 (C-2, C-3, C-4, C-5, C-6),
δ ) 137.47 (C-1), δ ) 157.14 (C6H5CH2OCONH-), δ ) 171.39
(N(CH2COOC(CH3)3)2), δ ) 173.10 ((H3C)3COCOCH). MS (FAB,
C30H48N2O8): M + H+ ) 565 m/z.
Synthesis of N-Succinyl-N-(N′-methyl-N′-(tert-butyloxycarbonyl)-
12-aminododecyl)octadecylamine (6). A solution of compound 5 (715
mg, 1.26 mmol), 400 µL of triethylamine, and 189 mg (1.89 mmol) of
succinic anhydride in 40 mL of absolute chloroform was stirred at room
temperature for 3 h. The organic phase was extracted twice with 1 N
NaOH and dried over anhydrous sodium sulfate. The solvent was
removed in a vacuum. Yield: 772 mg (1.16 mmol), 92%. TLC: Rf
) 0.53 in CHCl3/AcOH (25:1). 1H NMR (400 MHz, CDCl3): δ )
0.88 (t, 3 H, H3C(CH2)17N-), δ ) 1.25 (m, 46 H, H3C(CH2)15CH2-
CH2N(COCH2CH2COOH)CH2CH2(CH2)8-), δ ) 1.45 (s, 9 H, (H3C)3-
COCON-), δ ) 1.46-1.60 (m, 6 H, -CH2CH2N(COCH2CH2COOH)-
CH2CH2(CH2)8CH2CH2N(CH3)Boc), δ ) 2.68 (m, 4 H, NCOCH2-
CH2COOH), δ ) 3.18 (dd, 2 H, Boc-N(CH3)CH2-), δ ) 3.23, 3.31
(2 dd, 2*2 H, -CH2N(CO(CH2)2COOH)CH2-). 13C NMR (100 MHz,
CDCl3): δ ) 14.80 (H3C-),δ ) 23.38, 27.39, 27.59, 27.70, 28.75,
29.18, 29.51, 30.05, 30.24-30.38, 32.62 (H3C(CH2)16CH2N(CO(CH2)2-
COOH)CH2(CH2)10-), δ ) 34.74 (Boc-N(CH3)CH2-), δ ) 47.36
(-CH2N(CO-(CH2)2COOH)CH2-), δ ) 49.00 (N(COCH2CH2-
COOH), δ ) 49.50 (-CH2N(COCH2CH2COOH)CH2-), δ ) 79.79
((H3C)3COCON-), δ ) 156.60 ((H3C)3COCON), δ ) 172.86
(-CH2N(CO(CH2)2COOH)CH2-), δ ) 175.55 (-CH2N(CO(CH2)2-
COOH)CH2-). MS (FAB, C40H78N2O5): M + H+ ) 667 m/z.
Synthesis of N-Succinyl-N-(N′-methyl-12-aminododecyl)octade-
cylamine (7). The protected ω-amino acid 6 (772 mg, 1.16 mmol)
was dissolved in 44 mL of chloroform/trifluoroacetic acid (9:1) and
stirred at room temperature. After completion of the gas production
(12 h), the organic phase was extracted with 0.01 N NaOH until the
aqueous phase remained basic. The combined organic phases were
dried over sodium sulfate, and the solvent was removed in a vacuum.
Yield: 617 mg (1.09 mmol) of 7, 94%. TLC: Rf ) 0.31 in CHCl3/
MeOH (9:1). 1H NMR (400 MHz, CDCl3): δ ) 0.87 (t, 3 H, H3C-
(CH2)17N-), δ ) 1.25 (m, 46 H, H3C(CH2)15CH2CH2N(CO(CH2)2-
Synthesis of Nr,Nr-Bis[(tert-butyloxycarbonyl)methyl]-L-lysine
tert-Butyl Ester (11). The Z-protected amine 10 (614 mg, 1.09 mmol)
was dissolved in 50 mL of CHCl3/AcOH (25:1). After 15 mg of Pd/C
(5% Pd) was added, compound 10 was hydrogenated at room
temperature and normal pressure for 2 h. The catalyst was filtered
off, and the reaction mixture was extracted twice with 1 N NaOH. The
combined organic phases were dried over anhydrous sodium sulfate,
and the organic solvent was removed in a vacuum. Yield: 421 mg
(0.98 mmol) of 11, 90%. TLC: Rf ) 0.32 in CHCl3/MeOH (3:1). 1H
NMR (400 MHz, CDCl3): δ ) 1.43 (s, 18 H, ((H3C)3COCOCH2)2-
N-), δ ) 1.44 (s, 9H, (H3C)3COCOCH2CH-), δ ) 1.46, 1.63 (2*m,
COOH)CH2CH2(CH2)8-),
δ ) 1.48, 1.55 (2 q, 2*2 H,
CH2CH2N(CO(CH2)2COOH)CH2CH2-), δ ) 1.68 (q, 2 H, H3C+NH2-
CH2CH2-), δ ) 2.59 (m, 4 H, -CH2N(CO(CH2)2COOH)CH2-), δ

Molecular Recognition of Histidine-Tagged Molecules
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2757
2*2 H, H2N(CH2)2(CH2)2-), δ ) 1.73 (bm, 2 H, H2NCH2CH2-), δ )
2.68 (t, 2 H, H2NCH2-), δ ) 3.29 (t, 2 H, H2N(CH2)4CH-), δ )
3.45 (dd, 4 H, -N(CH2COOC(CH3)3)2). 13C NMR (100 MHz,
CDCl3): δ ) 23.88, 28.80-28.89, 31.18, 33.89 (H2NCH2(CH2)3-,
((CH3)3COCO-)3), δ ) 42.61 (H2NCH2-), δ ) 54.46 (-N(CH2-
COOC(CH3)3)2), δ ) 65.96 (H2N(CH2)4CH-), δ ) 81.31 (-N(CH2-
COOC(CH3)3)2), δ ) 81.66 (((CH3)3COCOCH)3), δ ) 171.39
(-N(CH2COOC(CH3)3)2), δ ) 173.10 ((CH3)3COCOCH). MS (FAB,
C22H42N2O6): M + H+ ) 431 m/z.
Synthesis of Nr,Nr-Bis[(tert-butyloxycarbonyl)methyl]-NE-[N-[N′-
methyl-N′-(7-nitro-2,1,3-benzoxadiazol-4-yl)-12-aminododecyl]oc-
tadecylamine]succinyl-L-lysine tert-Butyl Ester (12). Under the
exclusion of light, 715 mg (0.98 mmol) of compound 8 was dissolved
in 20 mL of absolute CH2Cl2 and 202 mg (0.98 mmol) of dicyclo-
hexylcarbodiimide, 113 mg (0.98 mmol) of N-hydroxysuccinimide and
15 mg of 4-(dimethylamino)pyridine in 10 mL of absolute acetone were
added. After 4 h of stirring at room temperature, the precipitated urea
was filtered off and 421 mg (0.98 mmol) of compound 11 and 390 µL
of triethylamine in 20 mL of absolute CH2Cl2 were added. The reaction
was stirred for 5 h. Thereafter the solvent was removed in a vacuum.
The product was recrystallized from acetone. Yield: 1005 mg (0.88
mmol) of 12, 90%. TLC: Rf ) 0.29 in CHCl3/MeOH (75:1). 1H NMR
(400 MHz, CDCl3): δ ) 0.89 (t, 3 H, H3C(CH2)17-), δ ) 1.18-1.40
(m, 46 H, H3C(CH2)15-, NBD-N(CH3)(CH2)2CH2)8-), δ ) 1.42 (s,
18 H, ((H3C)3COCOCH2)2N-), δ ) 1.44 (s, 9H, (H3C)3COCOCH2-
CH-), δ ) 1.45-1.67 (m, 10 H, -(CH2)CH2N(CO(CH2)2CONHCH2-
(CH2)2)(CH2)CH2-, NBD-N(CH3)CH2CH2-), δ ) 1.74 (q, 2 H,
-CO(CH2)2CONH(CH2)3(CH2)CH-), δ ) 2.50, 2.62 (2 t, 2*2 H,
-CO(CH2)2CONH-), δ ) 3.15-3.29 (m, 7 H, -CH2N(CO(CH2)2-
CONHCH2(CH2)3CH)CH2-), δ ) 3.43 (dd, 4 H, -N(CH2COOC-
(CH3)3)2), δ ) 3.47 (m, 3 H, NBD-N(CH3)-), δ ) 4.04 (m, 2 H, NBD-
N(CH3)CH2-), δ ) 6.06 (d, 1 H, H-6), δ ) 6.43 (b, 1 H,
-CO(CH2)2CONH-), δ ) 8.40 (d, 1 H, H-5). 13C NMR (100 MHz,
CDCl3): δ ) 14.74 (CH3-), δ ) 23.31, 23.84, 27.31-30.77, 32.35,
32.55, 39.98 (H3C(CH2)16CH2N(CO(CH2)2CONHCH2(CH2)3)CH2-
(CH2)10-, ((CH3)3COCO-)3), δ ) 42.15 (NBD-N(CH3)-), δ ) 46.86
(CO(CH2)2CONHCH2-), δ ) 48.61 (-CH2N(CO(CH2)2CONHCH2-
(CH2)3)CH2-), δ ) 54.49 (-N(CH2COOC(CH3)3)2), δ ) 56.63 (NBD-
N(CH3)CH2-), δ ) 65.76 (-CO(CH2)2CONH(CH2)3(CH2)CH-), δ
) 81.31 (-N(CH2COOC(CH3)3)2), δ ) 81.69 ((CH3)3COCOCH-), δ
) 101.55 (C-6), δ ) 122.74 (C-5), δ ) 136.00 (C-1), δ ) 145.16
(C-2), δ ) 145.52 (C-3), δ ) 146.09 (C-4), δ ) 171.34 (-N(CH2COOC-
(CH3)3)2), δ )171.99 (-CO(CH2)2CONH-), δ ) 173.00 ((H3C)3-
Figure 1. Chemical structure of the NBD-labeled chelator lipid 13
forming an octahedral complex (schematic) in the presence of nickel
14. Arrows indicate the free binding sites which can be occupied by
histidines. For comparison, the structure of the unlabeled chelator lipid
15 is illustrated.
CONH-), δ )173.88 (-CO(CH2)2CONH-), δ ) 175.49 (-+NH-
(CH2CO2H)2), δ ) 175.91 (HO2CCH).
Results and Discussion
Synthesis of Fluorescence-Labeled Chelator Lipids. A
variety of biochemical and biophysical methods are known to
study adsorption processes at lipid interfaces. However, in many
cases, it is not possible to distinguish between adsorption of
molecules to an interface on one hand and molecular recognition
of a molecule by a particular lipid in the assembly on the other
hand. Fluorescence energy transfer is well suited to monitor
pair formation due to the distance dependency of energy transfer
in the lower nanometer range. Thus, binding of an acceptor-
labeled biomolecule to a donor-labeled lipid can be followed
directly.
For our studies, we synthesized a metal-chelating lipid
carrying a fluorophore as a spectroscopic reporter (Figure 1).
To avoid interactions between the fluorophore and the metal-
chelating group we used NBD, which is supposed to stay in
the hydrophobic moiety of a lipid layer, whereas the NTA group
is accessible to water. We decided in favor of a NBD label
because of its spectral overlap and comparable fluorescence
lifetime to the acceptor fluorophore rhodamine which was used
to label the histidine-tagged molecule. In addition to its spectral
properties, NBD is highly suitable for fluorescence recovery
after photobleaching (FRAP) experiments to determine lateral
diffusion in lipid interfaces.
COCOCH-), δ ) 173.24 (-CO(CH2)2CONH-). MS (FAB, C63H112
-
N7O11): M + H+ ) 1143 m/z.
Synthesis of Nr,Nr-Bis(carboxymethyl)-NE-(N-(N′-methyl-N′-(7-
nitro-2,1,3-benzoxadiazol-4-yl)-12-aminododecyl)octadecylamine)-
succinyl-L-lysine (13). Under the exclusion of light, 1005 mg (0.88
mmol) of compound 12 was dissolved in 90 mL of chloroform/
trifluoroacetic acid (5:1). The reaction was stirred for 20 h at room
temperature and then extracted twice with 1 N NaOH. After recrys-
tallization from acetone, the product is purified by silica gel column
chromatography with CHCl3/MeOH/H2O (69:27:4) as the eluant.
Yield: 729 mg (0.75 mmol) of 13, 81%. TLC: Rf ) 0.38 in CHCl3/
MeOH/H2O (69:27:4). 1H NMR (400 MHz, CDCl3/CD3OD/F3CCOOD
(9:0.9:0.1): δ ) 0.73 (t, 3 H, H3C(CH2)17-), δ ) 1.05-1.56 (m, 54
H, H3C(CH2)16-, NBD-N(CH3)(CH2)2(CH2)9-, -CO(CH2)2CONHCH2-
(CH2)2-), δ ) 1.58-1.76 (m, 4 H, NBD-N(CH3)(CH2)CH2-,
-CO(CH2)2CONH(CH2)3CH2-), δ ) 2.34, 2.50 (2 t, 2*2 H,
-CO(CH2)2CONH-), δ ) 3.04 (m, 2 H, -CO(CH2)2CONHCH2-),
δ ) 3.12 (m, 4 H, -CH2N(CO(CH2)2CO)CH2-), δ ) 3.30 (t, 1 H,
-CO(CH2)2CONH(CH2)4CH-), δ ) 3.36 (m, 3 H, NBD-N(CH3)-),
δ ) 3.45 (m, 4 H, -+NH(CH2CO2H)2), δ ) 3.94 (m, 2 H, NBD-
N(CH3)CH2-), δ ) 6.03 (d, 1 H, H-6), δ ) 8.32 (d, 1 H, H-5), 13C
NMR (100 MHz, CDCl3/CD3OD/F3CCOOD (9:0.9:0.1): δ ) 14.34
(H3C-), δ ) 23.06, 24.14, 27.04-30.07, 31.60, 32.32, 39.29, 46.97,
48.31-50.12 ((H3C(CH2)17N(CO(CH2)2CONH(CH2)4)(CH2)11-, NBD-
N(CH3)-), δ ) 55.55 (-+NH(CH2CO2H)2), δ ) 56.51 (NBD-
N(CH3)CH2-), δ ) 66.48 (-CO(CH2)2CONH(CH2)4CH-), δ )
101.79 (C-6), δ ) 121.95 (C-5), δ ) 136.36 (C-1), δ ) 145.12 (C-2),
δ ) 145.37 (C-3), δ ) 146.32 (C-4), δ )172.70 (-CO(CH2)2-
The synthesis of lipid 13 is based on a modular system
whereby the lipid backbone 1-8 is synthesized separately from
the chelating headgroup 9-11 (Figure 2). Since the NBD
fluorophore had to be introduced in the hydrophobic moiety of
the lipid, we synthesized the asymmetrically substituted second-

2758 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Dorn et al.
Figure 2. Synthesis of the NBD-labeled metal-chelating lipid 13.
ary amine 5. This was accomplished via reduction of the
corresponding amide 3. Problems occurred during reduction
of the amide beside the carbamate of the Boc group in
compound 3. In contrast to amides substituted with shorter alkyl
chains,40 the reduction of the amide in compound 3 either failed
or resulted in complex mixtures of products using selective
reduction agents such as (Bu)4NBH4,41 NaBH3OCOCF3,42
LiBH4/TMSCl,43 LiAlH4-n(OMe)n,44 or BH3/THF.45 We finally
succeeded in quantitative reduction of the amide with the strong
reducing agent LiAlH4, but with the side effect of loosing the
Boc protecting group at the primary amine. Interestingly, the
release of the Boc group did not result in the primary amine as
expected. Instead, quantitative methylation of the primary amine
was observed. Although reductive alkylation of amines such
as the Leuckart-Wallach reaction are extensively described,40
methylation of primary amines via reduction of a Boc group
was not reported so far. The reduction of the Boc group was
verified by IR spectroscopy where no carbonyl valence vibration
either of an amide or a carbamate were detected (data not
shown). Reductive methylation of the primary amine was
(40) Glaser, H.; Mo¨ller, F.; Pieper, G.; Schro¨ter, R.; Spielberger, G.;
So¨ll, H. Methoden der organischen Chemie (Houben-Weyl); Georg Thieme
Verlag: Stuttgart, Germany, 1957; Vol. XI.
(41) Wakamatsu, T.; Inaki, H.; Ogawa, A.; Watanabe, M.; Ban Hetero-
cycles 1980, 14, 1437-1440.
(42) Umino, N.; Iwakuma, T.; Itho, N. Tetrahedron Lett. 1976, 33, 2875-
2876.
(43) Giannis, A.; Sandhoff, K. Angew. Chem. 1989, 101, 220-222.
(44) Brown, H. C.; Tsukamoto, A. J. Am. Chem. Soc. 1964, 86, 1089-
1095.
(45) Roeske, R. W.; Weitl, F. L.; Prasad, K. U.; Thompson, R. M. J.
Org. Chem. 1976, 41, 1260-1261.

Molecular Recognition of Histidine-Tagged Molecules
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2759
demonstrated: (i) by mass spectrometry, where the masses of
compounds 4-8 and 12 were increased by 14 amu (RNHCH3
1
instead of RNH2); (ii) by the high-resolution H NMR spectra
of 4-8, 12, and 13, where a singlet with an integration of three
protons was observed with a chemical shift depending on the
substitution of the amine (4: δ ) 2.43 ppm, no substituent; 5:
δ ) 2.83 ppm, Boc group; 7: δ ) 2.64 ppm, protonated; 8 and
12: δ ) 3.48 and 3.47 ppm, NBD); (iii) by thin-layer
chromatography, where the diamine 4 could not be stained with
fluorescamine, a specific reagent for primary amines.46 If the
primary amine group in amide 3 was deprotected prior to
reduction with LiAlH4, neither a shift in the masses of 4-8,
12, and 13 by 14 amu nor a singlet with an integration of three
protons in the 1H NMR spectra was observed (data not shown).
In this case the reduction product could be stained with
fluorescamine.
In the next step, we had to reintroduce regioselectively the
Boc group at the methyl-substituted amine in compound 4. This
was accomplished at -20 °C, whereas at higher temperatures
or with an excess of di-tert-butyl dicarbonate, both amines were
protected (data not shown). The following steps, coupling of
the succinyl spacer and the NBD group, are described in the
literature.47,48
The NTA derivative 11 was synthesized according to ref 23
but with a slight modification. The carboxy groups of NTA
were protected by tert-butyl esters which leads to a better
solubility of compound 11 in organic solutions and consequently
to better yields in the coupling of the lipid backbone 8 with the
chelator headgroup 11 (90%). In the final step, the tert-butyl
esters are cleaved with TFA and the product is isolated by silica
gel chromatography with an overall yield of 38%.
Pair Formation at the Monolayer Interface. The physical
behavior of lipid monolayers at the air-water interface can be
studied sensitively by film balance techniques. Molecular
packing of the lipids as well as docking of biomolecules onto
monolayers can be followed by changes in the area-pressure
isotherms in parallel to a variety of surface-sensitive techniques.
However, by using these methods, it is not possible to
distinguish between adsorption to a monolayer and formation
of a distinct lipid-protein complex. To follow the recognition
process between histidine-tagged biomolecules and metal-
chelating lipids, we studied the pair formation of a rhodamine-
labeled histidine peptide and the synthesized NBD-labeled
chelator lipid at the air-water interface by fluorescence
resonance energy transfer. For these studies we designed a
peptide composed of a C-terminal affinity tag of six histidines,
a hydrophilic and flexible glycine-serine spacer, and a N-
terminally coupled rhodamine fluorophore circumventing steric,
hydrophobic, or electrostatic contribution caused, e.g., by a
protein.
Figure 3. Fluorescence energy transfer at the air-water interface.
Fluorescence emission spectra of a monolayer of (a) SOPC and 3 mol
% of NBD-labeled Ni-NTA lipid 14 (solid line), after addition of 1
nmol (subphase concentration of 40 nM) of rhodamine-labeled histidine
peptide (dashed line) and after addition of 100 µM EDTA to the
subphase (dotted line), (b) SOPC and 3 mol % of NBD-labeled Ni-
NTA lipid 14 (solid line) and after addition of 1 nmol of unlabeled
histidine peptide (dashed line), (c) SOPC and 3 mol % of unlabeled
Ni-NTA lipid 15 and 3 mol % of NBD-labeled lipid 8 (solid line),
after the addition of 1 nmol of rhodamine-labeled histidine peptide
(dashed line) and after the addition of 100 µM EDTA to the subphase
(dotted line). All spectra were recorded at a lateral pressure of 20 (
0.5 mN/m and on 10 mM HEPES, 150 mM NaCl, pH 7.5, at 22 °C.
lateral diffusion the fluorescence is recovered immediately after
photobleaching of the fluorophore so that the NBD emission
remains constant during the recording of the emission spectra
(Figure 3a, solid line). At constant surface area, 1 nmol of
rhodamine-labeled histidine peptide was added, yielding a final
concentration of 40 nM in the subphase. Due to the high lateral
pressure and the low molar ratio of metal-chelating lipid, only
a slight increase in the lateral pressure of 0.5 mN/m was
observed over a period of 2 h, which is in the range of error. In
contrast to these minor thermodynamic changes of the lipid
monolayer, a drastic effect on the fluorescence emission spectra
was recorded after the addition of rhodamine-labeled histidine
peptide. The NBD emission intensity at 530 nm decreased about
50%, and a new fluorescence band at 570 nm according to the
emission maximum of rhodamine appeared (Figure 3a, dashed
line). This decrease in donor intensity to the expense of the
acceptor intensity provides strong evidence for a fluorescence
energy transfer caused by direct binding of the rhodamine-
labeled histidine peptide to the NBD-labeled chelator lipid. To
provide further proof for a molecular recognition event, we
added EDTA as a competitor for the nickel ion to the subphase.
After the addition of EDTA to a final concentration of 0.1 mM,
For the monolayer experiments, we chose SOPC as a matrix
lipid because of its high fluidity over a broad range of lateral
pressure and its length of alkyl chains matching best the alkyl
chains of the labeled chelator lipid. The SOPC monolayer was
doped with 3 mol % NBD-labeled Ni-NTA lipid 14, and an
area-pressure isotherm was recorded whereby a continuous
transition from the liquid-expanded to the liquid-condensed
phase was observed (data not shown). Up to 25 mN/m the
monolayer shows a homogeneous fluorescence. Because of high
(46) Udenfried, S.; Stein, S.; Bo¨hlen, P.; Dairman, W. Science 1972,
178, 871-872.
(47) Kung, V. T.; Reedemann, C. T. Biochim. Biophys. Acta 1986, 862,
435-439.
(48) Gosh, P. B.; Whitehouse, M. W. Biochemistry 1968, 108, 155-
159.

2760 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Dorn et al.
no rhodamine fluorescence was detected at the lipid interface
and 80% of the NBD emission was recovered with respect to
the beginning of the experiment (Figure 3a, dotted line). The
loss of fluorescence intensity can be explained by bleaching of
NBD at the air-water interface over a period of 6 h. In a
comparable set of experiments, 1 nmol of unlabeled histidine
peptide was injected underneath a monolayer of SOPC and 3
mol % NBD-labeled Ni-NTA lipid. Here again, the lateral
pressure increased about 0.5 mN/m over a period of 2 h. In
contrast to the experiment described before, the fluorescence
emission decreased only by 10% (Figure 3b, dashed line) which
is mainly due to NBD bleaching over the period of 2 h.
As final proof for pair formation between the histidine peptide
and the metal-chelating lipid, 1 nmol of rhodamine-labeled
peptide was injected underneath a monolayer identical to the
experiments described before but with one exception. In this
case, the fluorescence donor and the chelating group were not
located within the same lipid. Here, the SOPC monolayer was
doped with 3 mol % of unlabeled chelator lipid 15 and 3 mol
% of NBD lipid 8. After addition of the rhodamine-labeled
histidine peptide, the lateral pressure again increased about 0.5
mN/m. The NBD emission decreased about 25%, and a small
shoulder at 570 nm according to the rhodamine emission
maximum was observed (Figure 3c, dashed line). Addition of
EDTA to the subphase (0.1 mM) reduced the intensity at 570
nm (Figure 3c, dotted line) and increased the intensity at 530
nm only about 5%, indicating a much lower energy transfer
efficiency in comparison to the first set of experiments with
the fluorescent chelator lipid (for comparison see Figure 3a).
This can be explained by a larger average distance of donor
and acceptor since the chelator group and the fluorescence label
are not linked to the same lipid. Thus, direct docking of the
histidine peptide to the chelator lipid can be clearly distinguished
from adsorption of the histidine peptide to a chelator lipid
interface by the fluorescence energy transfer efficiency.
Figure 4. Fluorescence energy transfer in vesicle solution. Fluorescence
emission spectra of SOPC vesicles containing: (a) 3 mol % of NBD-
labeled Ni-NTA lipid 14, (b) 3 mol % of NBD-labeled Ni-NTA lipid
14, and 10 µM rhodamine-labeled histidine peptide, (c) 3 mol % of
unlabeled Ni-NTA lipid 15 and 10 µM rhodamine-labeled histidine
peptide. Spectrum c was subtracted from spectrum b to correct for direct
rhodamine excitation resulting in spectrum d. Spectrum e is calculated
by multiplying spectrum a by a factor of 0.54 to check for reabsorption
from unbound rhodamine.
chosen as matrix lipid. SOPC vesicles are unilamellar (100 nm
in diameter) and in a fluid phase at room temperature.49
Molecular docking of the histidine peptide to the chelator
lipid was demonstrated by a decrease in donor intensity and an
increase in acceptor intensity induced by fluorescence energy
transfer (Figure 4). To take the direct excitation of the
rhodamine into account, a spectrum of rhodamine-labeled
histidine peptide in the presence of unlabeled Ni-NTA lipid
vesicles (Figure 4, spectrum c) was recorded and subtracted from
the spectrum of NBD-labeled Ni-NTA lipid vesicles in the
presence of rhodamine-labeled histidine peptide (Figure 4,
spectrum b). The resulting difference spectrum (Figure 4,
spectrum d) confirms an efficient fluorescence energy transfer
from the donor-labeled chelator lipid to the acceptor-labeled
histidine peptide. A decrease of NBD intensity induced by
reabsorption from unbound rhodamine peptide can be excluded
due to the unchanged shape of the NBD emission spectrum up
to 540 nm. This is demonstrated by the comparison of spectrum
d with spectrum e, which was calculated from a spectrum of
NBD-labeled chelator lipid in the absence of rhodamine-labeled
peptide (Figure 4, spectrum a) by adjusting same intensities at
the NBD maximum.
We also checked the possibility that the binding of the
histidine peptide could induce phase separation or structural
reorganization of the fluorescence-labeled chelator lipid and
thereby change the spectral properties of the fluorophore. First,
we titrated SOPC vesicles containing 3 mol % of NBD-labeled
Ni-NTA lipid with up to 10 µM of unlabeled histidine peptide.
The intensity of the NBD remained unchanged within the range
of error (data not shown), indicating that quenching of the donor
emission by other effects besides fluorescence energy transfer
can be excluded. Second, we checked for self-quenching by
preparing SOPC vesicles which contain an increasing molar ratio
of the fluorescent chelator lipid. We observed a linear
dependency of the NBD fluorescence intensity up to 3 mol %
(Figure 5), indicating that no self-quenching of the NBD occurs
at molar ratios chosen in the experiments.
In all experiments the reversibility of the peptide binding
could be demonstrated by adding EDTA to the subphase as
competitor for the nickel. In both experiments energy transfer
was reduced after injection of EDTA (Figure 3a,c, dotted lines).
Although the rhodamine fluorescence vanished in Figure 3a,
rhodamine fluorescence at the monolayer can still be detected
in Figure 3c, indicating that there is still rhodamine-labeled
peptide immobilized at the monolayer. This may be due to the
higher amount of negatively charged lipids, 3 mol % of NTA
lipid and 3 mol % of NBD lipid 8, in the monolayer interacting
electrostatically with the histidine tag because at pH 7.5 a small
percentage of histidines are still positively charged. The
reversibility of the binding was also demonstrated by an increase
in NBD intensity after addition of EDTA. But in both
experiments only 80% of the NBD intensity was recovered with
respect to the intensity before addition of the rhodamine-labeled
peptide. This decrease of the NBD fluorescence in the
monolayer is most probably due to bleaching processes rather
than fluorescence energy transfer.
Molecular Docking at Vesicles. Vesicles as mimics for
cells, organelles, or biomembranes are well suited to study
molecular recognition processes at membranes. Furthermore,
in vesicle solution, the direct binding process of the rhodamine-
labeled peptide to the NBD-labeled chelator lipid can be
quantitatively analyzed by fluorescence energy transfer because
bleaching of the dyes was not observed and the absorbance of
the fluorophore can be measured precisely. For comparison
and because of its ideal mixing and phase behavior, SOPC was
To determine a binding constant of histidine-tagged biomol-
ecules for chelator lipids, we titrated SOPC vesicles containing
3 mol % of NBD-labeled chelator lipid with rhodamine-labeled
histidine peptide. Energy transfer efficiencies were calculated
(49) Cevc, G. Phospholipids Handbook; Marcel Dekker Inc.: New York,
1993.

Molecular Recognition of Histidine-Tagged Molecules
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2761
Figure 5. Fluorescence emission intensities at 530 nm of SOPC
vesicles containing various amounts of NBD-labeled Ni-NTA lipid
14. The linear regression curve is fitted to the first four values and
then extrapolated.
Figure 6. Binding constant of histidine peptide to vesicles doped with
Ni-NTA lipid 14 was determined by fluoresecence energy transfer.
Energy transfer efficiencies are plotted versus the concentration of
rhodamine-labeled histidine peptide (i) for specific binding in the
presence of Ni2+ (crosses) and (ii) for unspecific binding in the absence
of Ni2+ (stars). The chelator lipid concentration was 3.25 µM, and
measurements were performed in 10 mM HEPES, 150 mM NaCl, pH
7.5, at 22 °C. From the data a binding constant (Kd ) 3.0 ( 0.4 µM)
was calculated.
from the decrease of donor intensities because an increase in
acceptor intensity, especially at low acceptor concentrations,
was difficult to observe. The fluorescence emission spectra were
integrated from 490 to 540 nm to obtain a signal proportional
to the relative quantum yield of the donor which is unaffected
by the acceptor spectra. No change in the molar absorbance of
the donor during the titration of the peptide occurred (data not
shown). Therefore, we calculated the energy transfer efficien-
cies ET according to eq 4, where I and I0 correspond to the
Figure 7. Fluorescence correlation spectroscopy. (a) Time-dependent
fluorescence signal of rhodamine-labeled histidine peptides (15 nM)
was analyzed regarding to their different diffusion times (τ) in the free
or lipid-bound state. (b) Fluorescence autocorrelation functions G(t)
are shown in the presence of various amounts of SOPC vesicles
containing 3 mol % of the Ni-NTA lipid 14. The concentration of
metal-chelating lipid in the outer vesicle leaflet is given (0, 0.5, 1, 2.5,
5, 10, 20, and 40 µM). The autocorrelation functions are normalized
to an equal number of molecules in the confocal volume. (b) The
fraction of bound peptide is calculated from the autocorrelation functions
by a two-component fit and plotted versus the chelator lipid concentra-
tion which is accessible by the histidine peptide (outer vesicle leaflet).
Experiments were performed in the presence (crosses) and absence of
nickel ions (stars). The data are averaged from six measurements. From
the data a binding constant (Kd ) 4.3 ( 0.8 µM) was calculated.
ET ) 1 - (I/I0)
(4)
integrated fluorescence intensity of the donor in the presence
and absence of acceptor, respectively. Assuming a two-state
model of free and lipid-bound acceptor molecules, the energy
transfer efficiency is proportional to the amount of acceptor-
labeled histidine peptide bound to the donor-labeled NTA-
lipid.50 We plotted the energy transfer efficiencies against the
concentration of rhodamine-labeled histidine peptide (Figure 6,
crosses). The experimental data can be fitted by a Langmuir
isotherm (Figure 6, solid line)
efficiency at saturation. In the model, independent binding sites
are assumed which seem to be reasonable at receptor concentra-
tion of 3 mol %. From the fit a dissociation constant Kd of 3.0
( 0.4 µM was determined.
[A]ET∞
ET )
(5)
Kd + [A]
where [A] corresponds to the total concentration of the peptide,
Kd to the dissociation constant, and ET to the energy transfer
(50) Fo¨rster, T. Z. Naturforsch. 1949, 4a, 321-327.

2762 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Dorn et al.
Table 1. Diffusion Coefficients (D) Calculated from the Diffusion
Time (τ) through the Confocal Volume (1.5 fL)a
The fraction of lipid-bound peptide was calculated from a
two-component fit to the autocorrelation function of the time-
dependent fluorescence signal (eq 2). This fraction was
corrected for differences in the fluorescence quantum yield of
bound and free peptide using eq 6, where c corresponds to the
fluorophore
τ (ms)
D (10-10) (m2 s-1
)
rhodamine 6G (MW ) 478)
histidine peptide (MW ) 1976)
histidine peptide bound to vesicles
(100 nm L)
0.07
0.10
2.4
2.8
1.9
0.08
y
latex bead (100 nm L)39
0.04
c )
(6)
y + R2 - yR2
a Diffusion times t were calculated from a one- or two-component
fit to the autocorrelation function of the time-dependent fluorescence
signal.39 The concentration of the fluorophore was 15 nM in 10 mM
HEPES, 150 mM NaCl, pH 7.5, at 22 °C. The value of bound peptide
was measured at saturation (40 µM Ni-NTA lipid 14).
corrected fraction of lipid-bound peptide, y to the measured
fraction of lipid-bound peptide, and R to the ratio of counts per
molecules (cpm) in the bound and free state (in this experiment
R ) 0.58).39 The fraction of lipid-bound peptide was plotted
against the chelator lipid concentration (Figure 7b). Similar to
the fluorescence energy transfer measurements, the data can be
fitted by a Langmuir isotherm
To check for unspecific binding of the rhodamine-labeled
peptide, we performed the same experiment with SOPC vesicles
containing unloaded chelator lipid in the presence of 0.1 mM
EDTA. Even at peptide concentrations where saturation was
reached in the specific binding experiment (10 µM), an energy
transfer efficiency of only 3% was determined (Figure 6, stars).
In agreement with the monolayer experiments, an efficient
fluorescence energy transfer was observed due to the binding
of acceptor-labeled peptide to donor-labeled chelator lipid. The
histidine peptide serves as an ideal model for the vast variety
of histidine-tagged proteins since additional factors contributing
to the specific binding process are kept minimal. Therefore,
the binding constant (Kd ) 3 µM) represents a standard value
for the Ni-NTA system at self-assembled interfaces which can
be increased or decreased by additional steric, electrostatic, or
van der Waals contributions.
Fluorescence Correlation Spectroscopy. Alternatively, the
binding constant of histidine-tagged molecules to metal-chelating
lipids was examined by fluorescence correlation spectroscopy
(FCS).51 In contrast to the fluorescence energy transfer studies
where the lipid acted as a spectroscopic reporter of the binding
process, the rhodamine-labeled histidine peptide was used as a
reporter in the FCS measurements. Therefore, the binding of
the fast diffusing peptide to the slow diffusing SOPC vesicles
containing 3 mol % Ni-NTA lipid was followed by FCS. To
compare the results obtained from both methods, we used the
unlabeled chelator lipid 15, which has a structure very similar
to that of the NBD-labeled chelator lipid (Figure 1). For the
evaluation of the dissociation constant, we titrated 15 nM
rhodamine-labeled histidine peptide with vesicles containing
Ni-NTA lipid. From the fluorescence autocorrelation func-
tions, which were normalized to an equal number of molecules
in the confocal volume, an increase in the number of slow
diffusing particles (bound peptides) was detected to the expense
of the fast diffusing free peptides as the concentration of chelator
lipid was increased (Figure 7a). The diffusion times τ and the
diffusion coefficients D of free and lipid-bound molecules are
given in Table 1. The diffusion constant of the rhodamine-
labeled histidine peptide is slightly higher than for the standard
rhodamine 6G according to the higher molecular weight or more
precisely the larger hydrodynamic radius. Interestingly, the
diffusion constant of the peptide bound to the vesicle (100 nm
in diameter) is greater by a factor of 2 than the diffusion constant
of a latex bead with the same diameter. The apparent faster
diffusion of the vesicle bound fluorophore might be due to the
additional lateral diffusion of the fluorophore in plane of the
fluid membrane which does not occur with a fixed fluorophore
on a latex bead. Moreover, vesicles and latex beads might differ
in their hydrodynamic radii.
[A]R∞
R )
(7)
Kd + [A]
where R is the fraction of lipid-bound peptide, R∞ the fraction
of bound peptide at saturation, Kd the dissociation constant, and
[A] the concentration of NTA-lipid in the outer leaflet of the
vesicle assuming an equal distribution of NTA lipid between
the inner and outer leaflet. From the fit a dissociation constant
Kd of 4.3 ( 0.8 µM was obtained. To check for unspecific
binding, an identical experiment was performed with vesicles
containing unloaded chelator lipid in the presence of 100 µM
EDTA. Even at chelator lipid concentrations (40 µM) where
saturation of the specific binding was observed, only 3.7% of
the rhodamine-labeled peptide was bound to the vesicles.
The rather large difference in the diffusion coefficients of
the free and the vesicle-bound ligand makes fluorescence
correlation spectroscopy a well-suited method to study binding
of ligands to lipid vesicles in a homogeneous assay. Although
pair formation between histidine peptide and chelator lipid
cannot be distinguished from peptide adsorption to chelator lipid
containing vesicles, the very low unspecific binding of 3.7 %
strongly suggests that direct binding of the histidine peptide to
the chelator lipid assembled in SOPC vesicles occurs. The
determined dissociation constant Kd of 4.3 ( 0.8 µM is in good
agreement with the dissociation constant Kd of 3.0 ( 0.4 µM
determined by fluorescence energy transfer.
Conclusion
The concept of metal-chelating lipids is a versatile and
powerful method for the immobilization, orientation, and two-
dimensional crystallization of proteins at self-organizing inter-
faces. In this study, molecular recognition of a rhodamine-
labeled histidine peptide by a novel NBD-labeled chelator lipid
was demonstrated using fluorescence energy transfer at the
monolayer interface as well as in vesicle solution. Thereby we
were able to distinguish between a direct binding of the
histidine-tagged molecule to the chelator lipid and adsorption
onto the lipid interface due to the distance dependency of energy
transfer efficiency. The immobilization is highly specific,
leading to only 3% unspecific binding. An affinity constant of
3 µM was determined by fluorescence energy transfer and by
fluorescence correlation spectroscopy. In literature, binding of
histidine-tagged proteins to immobilized NTA- or IDA-metal
complexes is often characterized by a drastically higher affinity
than measured in this report. These high affinity constants were
not experimentally determined and mistaken for the formation
of the metal-chelate complex. The simple and effective one-
(51) Eigen, M.; Riegler, R. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5740-
5747.

Molecular Recognition of Histidine-Tagged Molecules
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2763
step purification of histidine-tagged proteins by immobilized
metal ion affinity chromatography (IMAC) is possibly due to
the number of theoretical plates during chromatography allowing
efficient rebinding after dissociation or from multivalent interac-
tions of the tagged protein with the affinity matrix. In this
report, the binding of histidine-tagged molecules to metal-
chelating lipids diluted into a self-assembled lipid matrix was
investigated. The affinity constant determined for the histidine
peptide by two independent methods is in agreement to the
binding constant of various histidine-tagged proteins analyzed
(I.T.D. and R.T., unpublished results). Furthermore, these
values are in the same range as the dissociation constant Kd of
1.5 µM reported for the binding of a histidine-tagged T-cell
receptor to Ni-NTA thiols self-assembled on gold.38 In another
publication, saturation for the immobilization of a histidine-
tagged polymerase to a Ni-NTA dextran surface was observed
at 2 µM.16 Despite the rather moderate dissociation constant,
kinetically stable immobilization of the proteins at the chelator
interfaces is reported for at least 60 min, demonstrating the
suitability of the chelator concept for bioanalytical studies.16,27,38
Detailed kinetic studies of the complex formation are under
investigation.
Acknowledgment. We thank Drs. Lutz Schmitt, Ulf Raedler,
Rudi Merkel, Joachim Ra¨dler, and Erich Sackmann for helpful
discussions and Drs. Wolfram Scha¨fer and Isolde Sonnenbichler
for providing the MS and NMR spectra. We are grateful to Dr.
B. Hecks (EVOTEC) and P. Bremer (Carl Zeiss Jena) for their
support and helpful comments concerning the FCS measure-
ments. This work was supported by the Deutsche Forschungs-
gemeinschaft and the Bundesministerium fu¨r Bildung und
Forschung.
JA9735620